<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html; charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#ffffff" text="#000000"> Dear all, <span class="Apple-style-span" style="border-collapse: separate; color: rgb(0, 0, 0); font-family: 'Times New Roman'; font-style: normal; font-variant: normal; font-weight: normal; letter-spacing: normal; line-height: normal; orphans: 2; text-indent: 0px; text-transform: none; white-space: normal; widows: 2; word-spacing: 0px; font-size: medium;">I think that the proposal of Jean-Michel of not making public the source codes of patented algorithms is the safest for IPOL, in front of hypothetical suits. Users can still run the algorithms on-line (are we sure that this does not infringe patent laws?) and they will have the guarantee by IPOL that what they run does indeed correspond to what is described in the web site. Moreover, I think that, as a matter of courtesy, the Editor of IPOL should send the authors of the patent a letter announcing them that there is an online version of their algorithm. I don't think we should ask for specific consent of the authors, since, in most cases, they don't even own the invention, their companies do. José Luis El 06/04/2011 7:42, Jean-Michel Morel escribió: <blockquote cite="mid:4D9BFD2A.6090001@cmla.ens-cachan.fr" type="cite"> Dear all, Should we  ask the advice of a patent attorney as suggests Nicolas?  It  is probably a good idea. However, unfortunately, I know some of these guys. They always give the most conservative advice. According to the CNRS patent attorney, to publish the NL-means code without the owner's consent IS an infringement because we help companies to infringe it, even though we mention that it is patented. We need a conservative solution showing our good faith. Here is what I propose: -systematically, the patented algorithms published in IPOL will be sent to their authors as referees. That way we will get a first hand information about whether and how the patent is enforced. We will have to follow the authors' advice. In case of touchy patent owners, we will only transmit the code to the referees under NDA. IPOL will therefore run online a code certified by referees, but which will not be public. In other terms IPOL will only publish (actually following its definition) "Algorithms", with as a bonus an implementation certified by referees, and running on line. The code of this implementation will be certified but not public.  The online facility will be installed in such a way (limited size, etc.) as to prevent any real or massive use.  Still, with all this, a patent owner could object that we are helping counterfactors who could do online reverse engineering with our code. But, at this point, I think our display of good faith will be solid, and it will all remain compatible with the IPOL definition that IPOL publishes algorithms and shows how they work. -Meanwhile, we never met the bad guys and we can continue disclosing the code. I agree with Pascal's proposal to add the text (enclosed below) to every patented algorithm. I do not think we can put any license on patented code without incurring in some risk.  Thus I agree that we should separate the patented part and put on it the warning proposed by Pascal. Nevertheless, it is not enough to write "for research and educational purposes": we must write "for non profit research and nonprofit educational purposes", because there is lucrative research and there is lucrative education. I disagree with Nicolas' proposal to put a BSD or other licenses, and invite users to decide by themselves if it is legal or not depending where they live. Putting a license at this point would be confusing. As a publisher we are supposed to know the law. It is our responsibility to avoid any ambiguity about rights. Best, JM Pascal Getreuer a écrit : <blockquote type="cite">_____________________________________________________  Copyright (c) 2011, Your Name Here  All rights reserved.  The source file "dmha.c" implements the algorithm covered by U.S. Patent  5629734.  For this reason, the source files "dmha.c", "dmha.h",  "dmhacli.c" and its binary program "dmha" may only be used for research  and educational purposes.  Redistribution or commercial use is not  allowed without the consent of the patent owner.  With the exception of the files mentioned above, redistribution and use in  source and binary forms, with or without modification, are permitted  provided that the following conditions are met:  (the rest of usual BSD license follows) _____________________________________________________ For files implementing the patent: "may only be used for research and educational purposes" and "redistribution or commercial use is not allowed without the consent of the patent owner." For all the other files: "With the exception of the files implementing the patent mentioned above, ..." (BSD/GPL license) So the BSD/GPL only applies to the files that are not covered by patents. Having this separate distinction is useful for example for the self-similar demosaicking code, where Hamilton-Adams is used as an initialization but actually the majority of the code is for the self-similar demosaicking algorithm itself.  The license specifies just the Hamilton-Adams part as "research and education use only, no redistribution" and everything else as GPL. Best, Pascal _______________________________________________ discuss mailing list <a class="moz-txt-link-abbreviated" href="mailto:discuss@list.ipol.im">discuss@list.ipol.im</a> <a class="moz-txt-link-freetext" href="http://tools.ipol.im/mailman/listinfo/discuss">http://tools.ipol.im/mailman/listinfo/discuss</a> </blockquote> _______________________________________________ discuss mailing list <a class="moz-txt-link-abbreviated" href="mailto:discuss@list.ipol.im">discuss@list.ipol.im</a> <a class="moz-txt-link-freetext" href="http://tools.ipol.im/mailman/listinfo/discuss">http://tools.ipol.im/mailman/listinfo/discuss</a> </blockquote> </body> </html>